Abstract
Access to accurate GSM power spectrograms in large cities can be of great interest to mobile network operators. Acquiring such information in urban settings is very challenging mainly due to the large area and high dynamics of environments. In this paper, we propose a novel scheme to tackle the large-scale mobile sensing problem. We first utilize an open source GSM project to collect RSSI values of all 194 GSM channels. Combining with moving vehicles, the sensing coverage can be enhanced. We leverage a commercial smartphone to perceive the moving direction and speed of the vehicle. With the motion information, the corresponding GSM power measures can be bound with the trajectory of the vehicle, forming the GSM power spectrograms. We have implemented the design of our system and conducted extensive on-road experiments. The results show the efficacy of the system.